AI RESEARCH
SCALE-LoRA: Auditing Post-Retrieval LoRA Composition with Residual Merging and View Reliability
arXiv CS.LG
•
ArXi:2605.01429v1 Announce Type: cross Libraries of Low-Rank Adaptation (LoRA) adapters are becoming a practical by-product of parameter-efficient adaptation. Once such adapters accumulate, a natural question is no longer how to train one adapter for one task, but how to reuse an open pool of adapters for a new task given only a small set. Prior work has shown that LoRA modules can be composed at the task level and dynamically selected at the instance level.